calculate overlap between the transcription data

[LinguList]

This will be useful for statistics, but also to see how much overlap we actually find.

[LinguList]

Interesting results (see statistics.md) and here (shared sounds per dataset):

TD1	TD2	sounds	percent
phoible	eurasian	619	0.28
phoible	pbase	528	0.29
phoible	lapsyd	540	0.33
eurasian	pbase	453	0.29
eurasian	lapsyd	417	0.29
pbase	lapsyd	359	0.34